skills.rest
quanganh208
quanganh208COMMUNITY
Explore all skills by this author →
GitHub

ck:context-engineering

Analyze LLM context utilization and degradation risks in agent systems.

1|Updated Mar 6, 2026
Search Tags:#multi-agent#compression#token-optimization#context-engineering#memory-systems#context-analysis
One-click install
npx skills add https://github.com/quanganh208/cookmate --skill ck-context-engineering-quanganh208
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ck:context-engineering
Source: https://github.com/quanganh208/cookmate/tree/main/.opencode/skills/context-engineering
Command: npx skills add https://github.com/quanganh208/cookmate --skill ck-context-engineering-quanganh208

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Large or long-running LLM sessions degrade performance, waste tokens, and cause failures like lost-in-middle and context poisoning; this Skill gives engineers actionable visibility and remediation steps to keep agent systems healthy and cost-effective.

Core Features & Use Cases

  • Context health analysis: Estimate token utilization, surface lost-in-middle risks, and compute composite health scores.
  • Compression & probe evaluation: Generate probes, measure compression ratio and quality, and recommend compaction strategies.
  • Runtime awareness & automation: Produce thresholds, usage warnings, and recommendations consumable by hooks or monitoring pipelines.
  • Use Case: Run the analyzer on a long-running multi-agent pipeline to find critical items buried in the middle, evaluate a compressed summary's fidelity, and get immediate compaction and artifact-tracking actions.

Quick Start

Run the context analyzer to assess token utilization and receive compaction and artifact-tracking recommendations for the current session.

Frequently Asked Questions about ck:context-engineering

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I detect lost-in-middle and context poisoning issues in long-running LLM sessions?▼

To detect lost-in-middle and context poisoning issues, analyze LLM context utilization to identify critical items buried in the middle and compute composite health scores for long-running sessions.

What is the best way to evaluate compression quality and measure token utilization for multi-agent pipelines?▼

Evaluating compression quality involves generating probes to measure compression ratio and fidelity, while token utilization metrics surface usage warnings and compaction recommendations for multi-agent pipelines.

How do I get actionable compaction recommendations for agent systems hitting token pressure?▼

You can get actionable compaction recommendations by running a context analyzer that estimates token pressure, tracks artifacts, and produces thresholds consumable by automation hooks.

Can I use context engineering analysis to monitor runtime token usage and trigger automated warnings?▼

Yes, context engineering analysis produces runtime usage warnings, thresholds, and recommendations that can be consumed by monitoring pipelines and automation hooks to manage token pressure.

When do I need context analysis for managing memory systems in large language models?▼

You need context analysis when large LLM sessions degrade performance and waste tokens, requiring visibility into context degradation risks and artifact tracking to keep systems cost-effective.

Does context engineering work for both development and production multi-agent scenarios?▼

Yes, context engineering analyzes and reports context degradation risks, artifact tracking, and token pressure across both development and production multi-agent scenarios.

Related Skills

Explore More Software Engineering Skills→
wangyouanwangyouan

context-compression

Compress large token contexts to minimize information loss in agent sessions.

Community
Advanced
ryasrkryasrk

benchmarking-dashboard

Monitor and visualize AI agent performance metrics with persistent dashboards.

Community
Intermediate
cryborg1211cryborg1211

vc:context-engineering

Diagnoses LLM context limits and reduces token usage for agent systems.

Community
Advanced
rohunvorarohunvora
5

context-compression

Compress long-running agent sessions into structured summaries with artifact indexes.

Community
Advanced